Comments on: Noodling with Atom/RDF

By: Morten Høybye Frederiksen

Morten Høybye Frederiksen — Tue, 11 Mar 2008 22:09:48 +0000

My current thinking, in this corner of the use case arena, is that a file in the cache represents an entry. So the filename is a key, which is generated off the entry id.

Whether that is good enough for real life, I’m not sure, but it’s good enough for Planet Venus, and since we’re downstream from that…

It does create issues, if the same entry is syndicated from multiple sources (it happens, think del.icio.us etc.), but I’m not sure there’s a better way to handle that case in general, differing properties and all.

By: danja

danja — Tue, 11 Mar 2008 09:31:17 +0000

Morten, cool – I’ll use yours instead

Re. Changesets – hmm, I must check, I was pretty sure some kind of (statement-replacing) update protocol is needed, but now I’m having my doubts, maybe David’s latest model avoids the need. (Henry figured out that the apparently simple Atom approach to identification actually uses combined inverse functional properties, which is a complication in the RDF world to say the least…).

By: Ian Davis

Ian Davis — Mon, 10 Mar 2008 23:29:20 +0000

Hi Morten. This theme is only temporary and we’ll certainly be including the author names in the replacement.

By: Morten Høybye Frederiksen

Morten Høybye Frederiksen — Mon, 10 Mar 2008 22:01:39 +0000

Heh.

Great minds think alike. Or something.

I was just yesterday thinking of ditching my WP-Venus in favour of something like this, most importantly powered by SPARQL.
Is the Changeset stuff necessary, you think?

PS: Any way you could add the author name or nick to the theme here, I had to check the feed to make sure it was you, Danny.